(2006) wind field model to simulate the maximum wind gust and the duration of strong winds [length of time that wind speeds are in excess of 20 m [s.sup.-1] (45 mph)] for each grid cell in the HOPM domain.
One method for quantifying the sensitivity of the HOPM predictions to variations in track and intensity is to calculate the coefficient of variation (coefficient of variation = standard deviation/ensemble mean) for each hurricane.
While it is instructive to look at the extremes from all 1,000 scenarios to demonstrate how variations in hurricane track and intensity can influence the HOPM, it is better to use the ensemble average (i.e., the average number of outages based on all 1,000 simulations) to predict the number of outages.
If the HOPM outages were normally distributed, this range could be calculated using mean and standard deviation from the 1,000 scenarios (e.g., lower bound = ensemble mean - 1 standard deviation; upper bound = ensemble mean + 1 standard deviation).
If a model such as HOPM provides only a point estimate, it is not providing the information needed by risk-averse decision makers, limiting their ability to make well-supported decisions about prestorm preparation activities.
1) Small errors in the official track and/or intensity forecast can lead to large errors in the resulting HOPM outage prediction.
As such, the MCWSP realizations used to run the HOPM in this study can be considered to be relatively plausible TC track and intensity scenarios.
3) The ensemble methodology used in this study yielded supplemental information about the HOPM prediction that may be useful to end users.